Mybatis 的一级缓存与二级缓存

缓存即存放在内存中的临时数据，利用缓存可以减少与数据库的交互次数、提高效率；经常查询且不经常改变的数据应该考虑缓存

Mybaits 支持的缓存有如下几个特点

• Mybatis 中默认定义了两级缓存：一级缓存和二级缓存
• 一级缓存：sqlsession 级别的缓存，也叫本地缓存，单个 sqlsession 期间生效
• 二级缓存：namespace 级别的缓存，也叫全局缓存，一个 xml 文件中的方法间生效
• 默认情况下，只有一级缓存开启，且不能关闭；二级缓存需要手动开启和配置
• 先说结论：Mybatis 在查询时先看二级缓存中有没有，再看一级缓存中有没有，最后才会查询数据库
• 工作中我们一般不会使用 mybatis 原生的缓存，而是使用 redis 作为缓存手段，这里主要是为了了解这种思想

一级缓存

测试案例 1：使用同一个 sqlSession 查询两次相同的记录

public void getUserById(){
    SqlSession sqlSession = MybatisUtil.getSqlSession();
    UserMapper mapper = sqlSession.getMapper(UserMapper.class);

    User user = mapper.getUserById(1);
    System.out.println(user);

    System.out.println("================================");

    User user1 = mapper.getUserById(1);
    System.out.println(user1);

    System.out.println(user == user1);

    sqlSession.close();
}

日志输出

Opening JDBC Connection
Created connection 1415157681.
Setting autocommit to false on JDBC Connection [com.mysql.jdbc.JDBC4Connection@545997b1]
==>  Preparing: select * from user where user_id = ? 
==> Parameters: 1(Integer)
<==    Columns: user_id, user_name, pwd, ctime
<==        Row: 1, 夏博阳, 000, 2021-11-03 14:07:53.0
<==      Total: 1
User(userId=1, userName=夏博阳, pwd=000, ctime=Wed Nov 03 14:07:53 CST 2021)
================================
User(userId=1, userName=夏博阳, pwd=000, ctime=Wed Nov 03 14:07:53 CST 2021)
true
Resetting autocommit to true on JDBC Connection [com.mysql.jdbc.JDBC4Connection@545997b1]
Closing JDBC Connection [com.mysql.jdbc.JDBC4Connection@545997b1]
Returned connection 1415157681 to pool.

日志分析：第一次查询时获取了 Connection 对象连接数据库进行查询，第二次没有获取；并且两次获取到的用户是同一个用户，通过 == 比较的结果为 true

测试案例 2：使用同一个 sqlSession 查询两次相同的记录，中间更新一下其他的用户

public void getUserById(){
    SqlSession sqlSession = MybatisUtil.getSqlSession();
    UserMapper mapper = sqlSession.getMapper(UserMapper.class);

    User user = mapper.getUserById(1);
    System.out.println(user);

    mapper.updateNameAndPwdById(new User(2, "小弟", "222", new Date()));
    System.out.println("================================");

    User user1 = mapper.getUserById(1);
    System.out.println(user1);

    System.out.println(user == user1);

    sqlSession.close();
}

日志输出

Opening JDBC Connection
Created connection 1415157681.
Setting autocommit to false on JDBC Connection [com.mysql.jdbc.JDBC4Connection@545997b1]
==>  Preparing: select * from user where user_id = ? 
==> Parameters: 1(Integer)
<==    Columns: user_id, user_name, pwd, ctime
<==        Row: 1, 夏博阳, 000, 2021-11-03 14:07:53.0
<==      Total: 1
User(userId=1, userName=夏博阳, pwd=000, ctime=Wed Nov 03 14:07:53 CST 2021)
==>  Preparing: update user set user_name = ?, pwd = ? where user_id = ? 
==> Parameters: 小弟(String), 222(String), 2(Integer)
<==    Updates: 1
================================
==>  Preparing: select * from user where user_id = ? 
==> Parameters: 1(Integer)
<==    Columns: user_id, user_name, pwd, ctime
<==        Row: 1, 夏博阳, 000, 2021-11-03 14:07:53.0
<==      Total: 1
User(userId=1, userName=夏博阳, pwd=000, ctime=Wed Nov 03 14:07:53 CST 2021)
false
Rolling back JDBC Connection [com.mysql.jdbc.JDBC4Connection@545997b1]
Resetting autocommit to true on JDBC Connection [com.mysql.jdbc.JDBC4Connection@545997b1]
Closing JDBC Connection [com.mysql.jdbc.JDBC4Connection@545997b1]
Returned connection 1415157681 to pool.

日志分析：第一次和第二次查询都获取了 Connection 连接数据库进行查询；两次获取到的用户不是同一个用户，通过 == 比较的结果为 false

一级缓存总结分析

• 增删改操作会使缓存失效，因为增删改有可能会改变数据；反之都会走缓存
• 手动调用缓存失效的方法

sqlSession.clearCache();

二级缓存

sqlsession 关闭时会将其中存储的一级缓存放入二级缓存中

二级缓存默认不会开启，通过如下步骤开启

1. 通过 setting 开启二级缓存

<settings>
    <setting name="cacheEnabled" value="true" />
</settings>

2. 在要使用二级缓存的 mapper.xml 中开启二级缓存

<!--在当前的 mapper.xml 文件中全局使用二级缓存-->
<cache />

<!--我们可以在 select 标签中通过 useCache 属性局部设置是否使用二级缓存-->
<select id="getUserById" resultType="user" useCache="false">
    select * from user where user_id = #{id}
</select>

<!--我们可以在增删改标签中通过 flushCache 属性局部设置执行 sql 后是否清空缓存-->
<update id="updateNameAndPwdById" parameterType="user" flushCache="false">
    update user set user_name = #{userName}, pwd = #{pwd} where user_id = #{userId}
</update>

测试案例 1：两次查询后一起 close sqlSession

public void getUserById(){
    SqlSession sqlSession = MybatisUtil.getSqlSession();
    SqlSession sqlSession1 = MybatisUtil.getSqlSession();
    UserMapper mapper = sqlSession.getMapper(UserMapper.class);
    UserMapper mapper1 = sqlSession1.getMapper(UserMapper.class);

    User user = mapper.getUserById(1);
    System.out.println(user);

    System.out.println("================================");

    User user1 = mapper1.getUserById(1);
    System.out.println(user1);

    System.out.println(user == user1);

    sqlSession.close();
    sqlSession1.close();
}

日志输出

Opening JDBC Connection
Created connection 1561408618.
Setting autocommit to false on JDBC Connection [com.mysql.jdbc.JDBC4Connection@5d11346a]
==>  Preparing: select * from user where user_id = ? 
==> Parameters: 1(Integer)
<==    Columns: user_id, user_name, pwd, ctime
<==        Row: 1, 夏博阳, 000, 2021-11-03 14:07:53.0
<==      Total: 1
User(userId=1, userName=夏博阳, pwd=000, ctime=Wed Nov 03 14:07:53 CST 2021)
================================
Cache Hit Ratio [com.xby.mapper.UserMapper]: 0.0
Opening JDBC Connection
Created connection 670971910.
Setting autocommit to false on JDBC Connection [com.mysql.jdbc.JDBC4Connection@27fe3806]
==>  Preparing: select * from user where user_id = ? 
==> Parameters: 1(Integer)
<==    Columns: user_id, user_name, pwd, ctime
<==        Row: 1, 夏博阳, 000, 2021-11-03 14:07:53.0
<==      Total: 1
User(userId=1, userName=夏博阳, pwd=000, ctime=Wed Nov 03 14:07:53 CST 2021)
false
Resetting autocommit to true on JDBC Connection [com.mysql.jdbc.JDBC4Connection@5d11346a]
Closing JDBC Connection [com.mysql.jdbc.JDBC4Connection@5d11346a]
Returned connection 1561408618 to pool.

日志分析：第一次和第二次查询都获取了 Connection 对象连接数据库进行查询；两次获取到的用户不是同一个用户，通过 == 比较的结果为 false

测试案例 2：第一次查询后先关闭该 sqlSession ，然后再进行第二次查询

public void getUserById(){
    SqlSession sqlSession = MybatisUtil.getSqlSession();
    SqlSession sqlSession1 = MybatisUtil.getSqlSession();
    UserMapper mapper = sqlSession.getMapper(UserMapper.class);
    UserMapper mapper1 = sqlSession1.getMapper(UserMapper.class);

    User user = mapper.getUserById(1);
    System.out.println(user);

    sqlSession.close();
    System.out.println("================================");

    User user1 = mapper1.getUserById(1);
    System.out.println(user1);

    System.out.println(user == user1);

    sqlSession1.close();
}

日志输出

Opening JDBC Connection
Created connection 1561408618.
Setting autocommit to false on JDBC Connection [com.mysql.jdbc.JDBC4Connection@5d11346a]
==>  Preparing: select * from user where user_id = ? 
==> Parameters: 1(Integer)
<==    Columns: user_id, user_name, pwd, ctime
<==        Row: 1, 夏博阳, 000, 2021-11-03 14:07:53.0
<==      Total: 1
User(userId=1, userName=夏博阳, pwd=000, ctime=Wed Nov 03 14:07:53 CST 2021)
Resetting autocommit to true on JDBC Connection [com.mysql.jdbc.JDBC4Connection@5d11346a]
Closing JDBC Connection [com.mysql.jdbc.JDBC4Connection@5d11346a]
Returned connection 1561408618 to pool.
================================
Cache Hit Ratio [com.xby.mapper.UserMapper]: 0.5
User(userId=1, userName=夏博阳, pwd=000, ctime=Wed Nov 03 14:07:53 CST 2021)
false

日志分析：第二次查询没获取 Connection 对象，是在缓存中获取的 User 对象；但两次获取到的用户不是同一个用户，通过 == 比较的结果为 false

二级缓存总结分析

• 只有当会话提交或者关闭时，才会将数据存储在缓存中
• 当 User 类没有继承序列化接口时，代码会报错，说明对象是序列化后缓存的，再次获取时是反序列化回来的新对象